Catalytic Carbon Oxidation in The Presence of ... - Wiley Online Library

Chem. Eng. Technol. 2009, 32, No. 12, 1859–1865

1859

May Issa1 Hakim Mahzoul1

Research Article

Alain Brillard1 Jean-Francois Brilhac1

Catalytic Carbon Oxidation in The Presence of Cerium Oxide: Experimental Study and Modeling of The Effect of Oxygen Concentration

1

Laboratoire Gestion des Risques et Environnement, University of Haute-Alsace, 25 rue de Chemnitz, 68200 Mulhouse, France.

The catalytic combustion of carbon black (CB) used as a model of diesel soot in tight contact with a commercial ceria (CeO2) was investigated. Oxygen mole fractions of 10, 5, and 1 % in the gas phase were tested in order to gain a better understanding of the redox properties of ceria and the mechanism of the catalytic oxidation of carbon black. Both isothermal and temperature programmed runs are performed to extract kinetic parameters, such as activation energy and reaction order with respect to oxygen partial pressure. The experimental data are used to propose a model of CB oxidation in the presence of CeO2 allowing the simulation of carbon oxidation. Keywords: Catalysis, Diesel soot combustion, Kinetics, Modelling, Oxidation. Received: October 4, 2008; accepted: July 6, 2009 DOI: 10.1002/ceat.200800509

1

Introduction

Anthropogenic activities like factories, motor vehicles, and construction activity are the main causes of air pollution. Among the various pollutants, soot particles are very harmful for human health and environment. Hence, stringent emission standards are imposed by the European Union in order to reduce soot emissions. Therefore, diesel particulate filters (DPFs) are now implemented on vehicles equipped with a diesel engine to comply with the stringent emission standards. They allow the trapping of soot in the gas exhaust pipe. The accumulated soot in the filter creates an undesirable backpressure in the gas exhaust line. Therefore, DPFs require to be periodically and/or continuously regenerated. As the temperature inside the DPF is ranging from 200 to 450 °C under normal driving conditions and soot oxidation by O2 occurs at a significant rate above 550 °C, it is crucial to use catalysts. They promote soot combustion at lower temperatures. They can be added in the form of a fuel borne catalyst or as a coating on the filter walls. In a catalyzed DPF, the nature of the contact between the soot particle and the catalyst particle is a key parameter in soot oxidation kinetics. Two types of contacts, a “loose” one and a “tight” one, were investigated in a previous study [2]. It was evidenced that “tight” contact mix-

– Correspondence: Prof. J.-F. Brilhac ([email protected]), Laboratoire Gestion des Risques et Environnement, University of Haute-Alsace, 25 rue de Chemnitz, 68200 Mulhouse, France.

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

tures have a higher oxidation rate at low temperatures than the “loose” contact ones [2]. Numerous catalysts were evaluated: (i) metal oxides [2, 3], (ii) molten salts [4, 5], (iii) perovskites, (iv) spinels [6], (v) mixture of metals, such as Cu-K-V-Cl [7], and Cu/K/Mo/Cl [8], (vi) precious metals like Pt supported on metal oxides [9]. Recently, it was reported that the activity of the catalyst for soot oxidation is related to its redox properties and its availability of surface oxygen [10–12]. The ability of cerium oxide to act as an oxidizing agent is due to its oxygen storage capacity and its redox properties (Ce4+/Ce3+) [10–12]. It exhibits large deviations from its CeO2 stoichiometric composition in a reducing atmosphere with a general formula, CeO2–x, with 0 ≤ x ≤ 0.5. At low concentration of oxygen, it is assumed that O2 concentration is not high enough to re-oxidize the cerium sites located at the surface of the catalyst. Among the mechanisms proposed for carbon oxidation by catalysts [3], the redox mechanism consists of carbon oxidation by lattice oxygen from the catalyst and re-oxidation of the catalyst by oxygen from the gas phase. De Soete [13] reported that adsorption of the oxidant on the metal sites is generally much faster than on the carbon sites. The author reported that the enhancement of the carbon combustion in the presence of a catalyst is due to the fact that carbon adsorbs metal bound oxygen atoms faster than molecular oxygen. Stratakis et al. [14] investigated the soot oxidation reaction in the presence of fuel additive containing an organometallic ceria compound. The kinetic parameters calculated from the TGA curves did not take into account the oxidation of reduced metallic sites.

http://www.cet-journal.com

1860


M. Issa et al.

In addition, the kinetics of the oxidation of additivated soot were determined by Millet et al. [15]. The authors reported that the re-oxidation reaction of reduced metallic sites was very fast and that any reduced additive was re-oxidized at once. The first objective of this paper is to determine the effect of the ratio, CB/CeO2, and the partial pressure of oxygen on the catalytic oxidation. The second objective is to propose a model simulating carbon black oxidation in the presence of a catalyst. The model is used to extract kinetic parameters for the catalytic oxidation rate with different molar oxygen fractions.

2

Experimental

2.1

Materials

A commercial ceria (CeO2) purchased from Rhodia Company (purity: 99.9 %) was used as a catalyst. A commercial carbon black (CB), Vulcan 6 from Cabot Company, was used as a model of diesel soot. Cerium oxide and carbon black were characterized by powder X-ray diffraction (XRD), using a Panalytical X’Pert Pro PMD diffractometer with Cu-Ka radiation (1.5405 A°) equipped with an Xcelerator detector. BET specific surface areas were measured with nitrogen as adsorbate at –196 °C, using Micromeritics ASAP 2010. The elemental analysis, BET, and XRD characterization results are shown in Tab. 1.

2.2

Various mixtures of carbon black and ceria oxide were prepared. Their composition varied from 50 to 95 % b.w. of catalyst and these samples (CB/CeO2) are referred to as 50/50, 25/75, and 5/95, respectively. The solids were grinded together in an agate mortar for 15 minutes to ensure a tight contact between them. The samples of CB/CeO2 were characterized by electron scanning microscopy (SEM) and the results are shown elsewhere [16]. SEM pictures allow to estimate the mean diameter of catalyst particles as reported in a previous paper [16].

Table 1. Elemental analysis, specific surface area, and XRD analysis for CB and cerium oxide. Elemental analysis [wt %]

BET surface area [m2/g]

XRD phase

107.9

amorphous

120

Cubic fluorite, Fm-3m

C: 95.3 H : 1.2 CB

O : 2.1 S: 1 N : < 0.3

CeO2

–

Activity Tests

CB oxidation with and without the presence of CeO2 was carried out in the fixed bed reactor set in a vertical electrical furnace. The sample was deposited on a quartz frit (16-mm internal diameter) set in a vertical fused silica tube. The composition of the injected gas mixture [x % O2, (100 – x) % N2] was determined using mass flow controllers. Different oxygen molar fractions were tested. The total gas flow rate was equal to 40 NL/h. Gas temperature was measured with a thermocouple (type K) located above the sample. Two types of experiments were performed: (i) temperature programmed oxidation (TPO) with a heating rate equal to 10 °C/min and (ii) isothermal oxidation. For the isothermal runs, the sample was heated from room temperature to the desired temperature under pure nitrogen and then the gas flow was switched to the oxidant mixture. CO and CO2 concentrations at the reactor outflow were measured by an infrared analyzer (COSMA dm CRISTAL 300). The oxidation rate ( ) was calculated from dt CO and CO2 emissions using the following relation: dm dt


XCO XCO2 Q MC

(1)

where XCO and XCO2 are the measured molar fraction in the gas phase, Q is the molar flow of the used gas (mol·s–1) and MC is the molar mass of carbon (kg·mol–1). The conversion rate of carbon black, ft , is calculated using the following equation: ft

Sample Preparation

Material

2.3

m0

mt m0

(2)

where m0 and m(t) are the initial mass and the remaining mass of CB in the sample at time t, respectively. The initiation temperature (Ti) is that at which 5 % of carbon is consumed. The temperature at which the conversion rate is equal to 65 % is denoted Tmax.

3

Results and Discussion

3.1

Effect of The Carbon/Catalyst Ratio

The oxidation rate versus temperature is plotted for all investigated samples tested in Fig. 1. In the absence of CeO2, carbon black oxidation starts around 560 °C and its rate reaches a maximum near 670 °C. The effect of the presence of CeO2 on carbon black oxidation is clearly observed with sample 50/50. Carbon black oxidation starts at a lower temperature (380 °C) with a maximum value of the oxidation rate at 505 °C (compared to 670 °C with CB alone). In the case of sample 25/75, carbon black oxidation starts at 360 °C and the oxidation rate reaches a maximum around 465 °C. An increase of the mass fraction of CeO2 in the mixture, 5/95, strongly enhances the rate of CB oxidation. In that case, carbon black oxidation begins at 350 °C and then rises progressively to reach a maximum value around 425 °C. The rate of CB oxidation in sample 5/95 is faster than those obtained with samples 50/50 and 25/75. These results show the clear effect of the catalyst/carbon black ratio on the oxidation



Catalytic Carbon Oxidation

Vspe

1.40E-08 1.20E-08 Oxidation rate (kg/s)

5/95

1861

1 dmt mt dt

(3)

The specific reaction rate increases during the initial stage of the reaction (until 4 % of conversion) and then decreases slightly with the extent of the reaction. There is a significant effect of the O2 concentration on the specific oxidation rate. It is also shown that catalyzed CB oxidation depends on the oxygen partial pressure.

CB

1.00E-08 50/50 8.00E-09 6.00E-09 25/75 4.00E-09

3.3 2.00E-09 0.00E+00 300

400

500

600

700

Temperature (°C)

The following equation was used to determine the kinetic parameters for the catalyzed carbon oxidation by assuming that the reaction order with respect to the remaining mass of carbon black is equal to 1: dm t n k T m t PO 2 dt

Figure 1. Oxidation rate of carbon black versus temperature for all samples.

(4)

8.00E-09 10%O2 7.00E-09 Oxidation rate (kg/s)

process. This is attributed to the surface of contact developed between the two solids. It is proportional to the catalyst/carbon ratio. With the increase of the catalyst fraction in the mixture, carbon black is totally surrounded by the catalyst. Indeed, the procedure used to the preparation of the samples permits to obtain a tight contact between carbon black and cerium oxide. As already underlined in a previous study [1], the tightness of the contact between the two solids is a critical factor for the enhancement of CB oxidation.

Reaction Order with respect to O2

5%O2

6.00E-09 5.00E-09

1%O2

4.00E-09 3.00E-09

180 ppm O2

2.00E-09 1.00E-09 0.00E+00 300

Effect of O2 Concentration on The Catalytized Carbon Oxidation

3.2.1 TPO Experiments The influence of O2 concentration on carbon black oxidation in the presence of cerium oxide was evaluated. The oxidation rate versus temperature for sample 50/50 is depicted in Fig. 2. A similar value for Ti was obtained for O2 concentration equal to 5 % and 10 % O2. Ti increases from 376 °C (for 10 % O2) to 400 °C (for 1 % O2) and Tmax increases from 500 to 540 °C. The oxidation rate of carbon black becomes slower in the presence of 180 ppm of oxygen. The oxidation rate curves versus T exhibits a very different profile in this case. CB oxidation occurs at higher temperatures compared to the other samples.

3.2.2 Isothermal Experiments CB oxidation in sample 25/75 was investigated at 424 °C with three different molar fractions of O2. The specific oxidation rate of carbon versus the conversion rate is plotted in Fig. 3. The specific reaction rate is calculated using the following equation:


350

400

450

500

550

600

650

700

Temperature (°C)

Figure 2. Oxidation rate of carbon black versus temperature for all samples.

3.50E-03 Specific oxidation rate (kg/s)/(kg)

3.2

3.00E-03 2.50E-03 2.00E-03

10%O2 5%O2

1.50E-03

1%O2

1.00E-03 5.00E-04 0.00E+00 0

20

40 60 % conversion

80

100

Figure 3. Specific reaction rate of carbon black in sample 25CB/75CeO2 as a function of carbon conversion rate for T = 424°C.


1862


M. Issa et al.

where k(T) and m(t) are the constant rate (Pa–1s–1) and the remaining mass of CB in the sample (kg) at time t, respectively. PO2 is the partial pressure of oxygen (Pa); n is the order of catalyzed reaction with respect to oxygen. n was determined from the isothermal runs performed with samples 50/50, 25/75, and 5/95. Its value is extracted from the ln(Vspe) versus ln(% O2) plot for a considered rate of conversion, ft , as depicted in Fig. 4. The slope of the line is equal to the reaction order with respect to the oxygen partial pressure. The values of n obtained for two samples (25/75 and 5/95) are listed in Tab. 2. It is seen that n varies slightly as a function of the conversion rate, ft . Fino et al. [17] have pointed out that oxygen reaction order does not depend on the amount of carbon available. An average value of n is obtained for all samples and conversion rates. It is close to 0.2. To our knowledge, no study has been reported in the literature to determine the order reaction with respect to oxygen with CeO2. An order value equal to 0.3 was found by Fino et al. [17] for the reaction catalyzed by a perovskite catalyst La1.8K0.2Cu0.9V0.1O4. The reaction order with oxygen for the uncatalyzed reaction (not shown here) is equal to 1. This value indicates that the rate of the reaction catalyzed by CeO2 depends less on the partial pressure of O2 in the phase gas. This may be attributed to the occurrence of oxygenated species and/or oxygen active species generated by the catalyst. Such species are also suggested by Fino et al. [17]. These authors indicate that an order value lower than 0.5 suggests that the oxygen intermediate species active in the oxidation reaction may be atomic oxygen generated by dissociative chemisorption over the catalyst surface.

Table 2. Values of the order reaction in O2 for two samples, 25/ 75 and 5/95, as a function of the conversion rate, ft . ft

n (25/75)

n (5/95)

10

0.1943

0.1969

15

0.2071

0.1957

20

0.2025

0.1978

25

0.197

0.1986

30

0.1943

0.1994

35

0.1908

0.1961

40

0.1874

0.1873

45

0.1871

0.1929

50

0.1792

0.1908

[%]

Ln (Vspec)

the mechanisms of the two reactions are not identical. Catalyzed carbon oxidation depends essentially on the interactions between the carbon black and the catalyst. Uncatalyzed carbon oxidation involves the dissociative chemisorption of molecular oxygen on free carbon sites. The dissociative chemisorption is limited to the edges of the crystallite of the carbon surface, i.e., the active sites [18]. Accordingly, the reaction order with respect to O2 depends on the number of carbon sites available for oxygen adsorption. It is well established that the first step of uncatalyzed carbon oxidation is the dissociative chemisorption of O2 on the carbon surface, leading to the formation of oxygen complexes. In contrast, in the presence of the catalyst, the adsorption rate of oxygen is much larger over the catalyst than on carbon 3.4 Discussion surface as reported by De Soete [13]. Hence, the catalytized carbon oxidation would be an indirect reaction. This fact was The difference between the values of reaction order obtained confirmed also by Bueno-Lopez et al. [19]. Indeed, these for the uncatalyzed and catalyzed CB oxidation indicates that authors underline the fact that the carbon oxidation rate depends on oxygen species present on the Ln(%O2) -6.5 surface of CeO2. 0 0.5 1 1.5 2 2.5 Accordingly, the sequential steps of the oxidation reaction of carbon in the presence of cerium -6.6 oxide are the following: Slope of the line = 0.1943 – Dissociative chemisorption of molecular oxygen -6.7 over the catalyst surface; – Oxygen activation on the catalyst surface; -6.8 – Transfer of active oxygen species from the catalyst to carbon black sites in tight contact with -6.9 CeO2. Therefore, the oxygen species involved in the ceria-catalyzed carbon reaction have two origins: -7 gaseous oxygen and oxygen species released from CeO2. It is indeed well known that CeO2 has a high -7.1 oxygen storage capacity. Consequently, the involvement of cerium oxide in the carbon oxidation -7.2 reaction results in a lower oxidation state and/or to a non-stoichiometric composition of the oxide Figure 4. Determination of reaction order in oxygen for the catalyzed carbon oxi(Ce2O3 or CeO2-x with 0 ≤ x ≤ 0.5). dation for sample 25/75.





A redox mechanism [12] allows the description of the oxidation of carbon black in the presence of a metal oxide MO2 (CeO2). It is illustrated in Fig. 5. It accounts for two reactions: – oxidation of carbon by the metal oxide under its oxidized form (CeO2): (–C) +2 CeO2 → CO2 + 2 Ce2O3 – re-oxidation of the reduced metallic sites with oxygen: Ce2O3 + 0.5 O2 → 2 CeO2 Reduced ceria results from the removal of oxygen from the CeO2 lattice, which generates oxygen vacant sites. The presence of oxygen in the gas phase is essential to fill oxygen vacancies and to promote the re-oxidation of reduced ceria, even though migration and/or diffusion of oxygen from the bulk to the surface vacancies occur. Balducci et al. [20] reported that facile oxygen diffusion takes place through the bulk for cubic CeO2-ZrO2, and that the high mobility of oxygen vacancies through the bulk and on the surface will assist the Ce4+/Ce3+ redox cycle.

4

Modeling

A one-dimensional model was developed for the simulation of CB oxidation in the presence of CeO2 and is described elsewhere [1]. It was used to simulate both isothermal and temperature programmed oxidations. The redox mechanism [12] is considered to describe the oxidation of carbon black in the presence of a metal oxide MO2 (CeO2). To our knowledge, no investigation concerning the kinetic parameters of the reduction and re-oxidation of cerium oxide CO CO2

CO2

O2

1863

in the presence of soot has been published so far. As mentioned before, the oxidation rate of Ce2O3 is generally considered to be very fast as compared to the reduction rate of CeO2 during CB oxidation [13, 15]. Hence, cerium is assumed to be mainly in its fully oxidized CeO2 form. The dependance of the carbon oxidation rate on the contact area between carbon and ceria particles was taken into account in a previous work [1] and the main points are summarized here below. The rate of carbon black oxidation is assumed to be proportional to the contact area exposed by the catalyst (Ai) and to the catalyst/carbon black ratio in the tested sample (s) at t = 0. It may be expressed for a given class of catalyst i as follows: dmi n t kT Ai s mi t PO 2 dt

(5)

with mi(t): mass of carbon remaining in contact with the family i of the catalyst in the sample (kg); k(T): rate constant of carbon oxidation in the presence of the catalyst (kg m–2 s–1); Ai: represents the surface developed by CeO2 particles per unit of mass of catalyst. This surface can be expressed as: Ai

Ni 4P R2i Ni

4 3

P R3i q

3 Ri q

(4)

where Ri is the catalyst radius (m) and q is the catalyst density (kg/m3). The gas temperature is Tg(t) = T0 + C × t, where T0 is the initial temperature and C is the heating rate (K.s–1). A thermal equilibrium between the sand and the gas phase is assumed (T(t) = Ig(t)). The rate constant, assumed to follow a first-order Arrhenius-type expression, is expressed as: Ea kT k0 exp (6) RT The overall rate of CB oxidation is given by: X dm t i dt

CeO2

CO CO2

CO2

O2 O

O

CeO2

Figure 5. Scheme of catalyzed carbon oxidation by CeO2 (& : represents an oxygen vacant site; the dashed arrow describes oxygen migration).


dmi t dt

(7)

The above equations are numerically solved through Euler’s method. A numerical discretization as a function of the size of catalyst particles and time is used. Ea is taken equal to 124 kJ mol–1 as described in [1] and this value is used as an input of the model to simulate isothermal oxidation. Shangguan et al. [21] have shown that Ea of the catalyzed reaction does not change with a decrease of the oxygen partial pressure. Therefore, the same activation energy is used to simulate carbon oxidation rate with different oxygen concentrations. k0 is obtained by fitting and is found equal to 45 kgCBm–2s–1. The results of the model predictions and the experimental values are compared for 2 oxygen concentrations (10 % O2 and 5 % O2) in Fig. 6a) and Fig. 6b), respectively. It shows that the model is able to predict, with good agreement, the oxidation behavior of CB for sample 25/75 under isothermal conditions.


1864 (a)


M. Issa et al.

3.50E-09 dm/dt (25CB/75CeO2)(10%O2) 3.00E-09

Oxidation rate (kg/s)

simulation (10%O2) 2.50E-09 2.00E-09 1.50E-09

tion is larger than 5 % in the gas phase. The model developed previously in [1] was refined by taking into account the partial pressure of oxygen. This model enables one to satisfactorily predict both temperature programmed oxidation and isothermal experiments. Kinetic parameters for the catalyzed carbon oxidation were obtained.

Acknowledgements

1.00E-09

The authors gratefully thank REALISE (Réseau Alsace de Laboratoires en Ingénierie et Sciences pour l’Environnement) for the financial support.

5.00E-10 0.00E+00 0

500

1000

1500

2000

The authors have declared no conflict of interest.

Time (s) (b)

References

3.00E-09 dm/dt (25CB/75CeO2)(5%O2)


2.50E-09

simulation (5%O2)

2.00E-09 1.50E-09 1.00E-09 5.00E-10 0.00E+00 0

500

1000

1500

2000

Time (s)

Figure 6. Oxidation rate of carbon black in sample 25/75 as a function of time for Ea = 124 kJ mol–1, k0 = 45 kgCBm–2s–1, and T = 424°C for 10, 5 % O2, respectively.

5

Conclusion

The aim of this study was to investigate carbon black oxidation in the presence of CeO2 and to develop a model allowing the simulation of carbon oxidation based on the experimental results. A dependence of the catalytized carbon oxidation rate on the oxygen concentration is evidenced. The rate of CB oxidation is higher when oxygen concentra-


dm/dt (50CB/50CeO2) (10%O2) dm/dt (50CB/50CeO2) (5%O2) dm/dt (50CB/50CeO2) (1%O2) simulation (10%O2)

8.00E-09 7.00E-09


The model and the kinetic parameters defined previously were then used to simulate the temperature programmed oxidation (TPO) experiments for sample 50/50. The comparison between the calculated values and experimental ones is shown in Fig. 7. It is seen that the model enables the simulation of the experimental results for all tested molar oxygen fractions (10, 5, and 1 % O2) for sample 50/50, with a good agreement.

[1] M. Issa, C. Petit, A. Brillard, J.-F. Brilhac, SAE technical paper series 2007-24-0091. [2] J. P. A. Neeft, M. Makkee, J. A. Moulijn, Chem. Eng. J. 1996, 64, 295. [3] G. Mul, F. Kapteijn, C. Doornkamp, J. Moulijn, J. Catal. 1998, 179, 258. [4] B. A. A. L. van Setten, J. M. Schouten, M. Makkee, J. A. Moulijn, Appl. Catal. B 2000, 28, 253. [5] A. Setiabudi, N. K. Allart, M. Makkee, J. A. Moulijn, Appl. Catal. B 2005, 60, 241. [6] D. Fino, N. Russo, G. Saracco, V. Specchia, J. Catal. 2003, 217, 367. [7] P. Ciambellli, V. Palma, P. Russo, S. Vaccaro, J. Mol. Catal. A: Chem. 2003, 204–205, 673. [8] G. Mul et al., Appl. Catal. B 1995, 6, 339. [9] J. Oi-Uchisawa et al., Appl. Catal. B 1999, 21, 9. [10] J. Kaspar, P. Fornasiero, J. Solid State Chem. 2003, 171, 19.

6.00E-09 5.00E-09

simulation (5%O2) simulation (1%O2)

4.00E-09 3.00E-09 2.00E-09 1.00E-09 0.00E+00 350

400

450

500

550

600

650

700

Temperature (°C)

Figure 7. Oxidation rate of carbon black as a function of temperature for Ea = 124 kJ·mol–1 and k0 = 45 kgCBm–2s–1.



[11] E. Aneggi et al., J. Alloys Compd. 2006, 408–412, 1096. [12] A. Bueno-Lopez, K. Krishna, M. Makkee, J. A. Moulijn, J. Catal. 2005, 230, 237. [13] G. De Soete, in Western States section meeting of the Combustion Institute, Salt Lake City, 1988. [14] G. A. Stratakis, A. M. Stamatelos, Comb. Flame 2003, 132, 157. [15] C. N. Millet, P. Ménégazzi, B. Martin, H. Colas, Oil Gas Sci. Technol. Rev. IFP 2003, 58, 151.



1865

[16] M. Issa, C. Petit, A. Brillard, J.-F. Brilhac, Fuel 2008, 87, 740. [17] D. Fino, P. Fino, G. Saracco, V. Specchia, Appl. Catal. B: Environm. 2003, 43, 243. [18] J. P. A Neeft, Ph.D. Thesis, Delft, The Netherlands 1995. [19] A. Bueno-López et al., Catal. Today 2007, 121, 237. [20] G. Balducci et al., J. Phys. Chem. B 1997, 101, 1750. [21] W. F. Shangguan, Y. Teraoka, S. Kagaw, App. Catal. B: Environm. 1997, 12, 237.
